Question about H1/H4 candle finalization and timestamp consistency
by Mahmoud
EASE: TickerAll is now live! Start free → × TICKER ALL! ≡ Contact Get in touch Questions, help, or a feature idea? Chat with us live, post on our public issues board, or email — a real person reads every message. Live chat Fastest way to reach us — chat opens right here on the page. A real person, usually within minutes. Issues & requests Report a bug, request a feature, upvote what matters, and follow what we're building — all public. Or email us General & sales hello@tickerall.com Product questions, Enterprise plans, partnerships — or anything else. Support support@tickerall.com Account or technical help. Include your account email so we can find you fast. Billing billing@tickerall.com Invoices, plan changes, and payment questions. Find us Prefer to dig in first? The API docs cover setup, endpoints, and examples, and the status page shows live system health. © 2026 TickerAll · hello@tickerall.com Docs Live Sign up ▲ Contact · Ticker All! Message us A real person reads every message. × Hi, I’m investigating candle-data consistency for XAUUSD_ using the candle API. I’m seeing several behaviors that I need clarification on: For H1 and H4 candles, can OHLC values be revised after the candle’s nominal close time? If yes, how long after close should a candle be considered final/immutable? I observed the same raw timestamp/epoch returned later with different OHLC values. Is this expected behavior? I also observed timestamps occasionally shifting by 1 second between identical fixed requests. For H4, I have seen candle grids such as 00/04/08 UTC in historical data, while another fixed request returned 01/05 UTC candles with matching OHLC shifted by 3 hours. Could you please confirm: Is the candle timestamp always the UTC bar OPEN time? What is the official H1/H4 candle finalization policy? Can historical closed candles be revised? What determines the H4 candle boundary/grid? Can the from / to request boundaries change H4 aggregation? Is a ±1 second timestamp variation expected? I’m using fixed ISO-8601 UTC Z request bounds and preserving the raw provider timestamps without modification. This information is important because I need to know when a closed candle is safe to treat as finalized.